Drought-induced reduction in methane fluxes and its hydrothermal sensitivity in alpine peatland

Site description

The experiment was conducted in Zoige county in the eastern Tibetan Plateau (33.79°N, 102.95°E) at an altitude of 3,430 m (Fig. 1A). The mean annual temperature is 1.1 °C, and mean annual precipitation is 648.5 mm, with 80% falling during the growing season from June to September. The mean monthly temperature ranges from 1 °C (January) to 11 °C (July). The experiment was established in a frigid temperate zone steppe dominated by herbaceous marshes and composed mainly of Carex meyeriana, Carex muliensis, and Kobresia tibetica. The main soil type was marshy peat, with the soil pH is between 6.8–7.2 in localized areas (Zhou et al., 2015). The depth of peat in the vertical profile of this site is in general 1.2 m. Field experiments were approved by the Institute of Wetland Research.

Experiment design and data collection

Based on the local rainfall data for the past 50 years, we defined daily rainfall ≥ three mm as ecologically effective precipitation (Hao et al., 2012). During the flourishing period of the growth season, we selected 32 days as the duration of non-ecologically effective precipitation (drought days) and simulated extreme drought over this period of plant growth (Wang et al., 2007). The area of the plot was 20 m × 20 m, and extreme drought treatment (D) and control treatments (CK) were set up, independently, with each treatment consisting of three (2 m ×2 m) repetition plots (Fig. 1B). We buried iron sheets in the soil about 1 m deep around each treatment to prevent the lateral flow of soil water. A stainless-steel base (50 cm ×50 cm ×20 cm) was placed at the sampling point and inserted into the ground at a depth of 10 cm. Before each measurement, we filled the groove of stainless steel with water to ensure the airtightness of the measurement (Fig. 1C). For the extreme drought treatment, we used a magnesium-aluminum alloy shelter (length × width × height; 2.5 m × 2.5 m × 1.8 m) to simulate drought, and the light transmittance of the shelter was more than 90%. The gas in the controlled plot was monitored under natural conditions.

A fast greenhouse gas analyzer (DLT-100, Los Gatos Research, USA) was used to monitor CH₄ fluxes, at a data acquisition frequency of 1 Hz. A TZS-5X thermometer was used to monitor the air temperature (Ta) and soil temperature (Ts), and a TDR 300 was used to measure the SWC. A box (50 cm ×50 cm) was connected with the fast greenhouse gas analyzer. There were two small holes two cm in diameter at the top of the box, which were closed with rubber plugs. There was a small hole in each rubber plug for the insertion of two gas conduits (intake pipe and outlet pipe) with a length of 20 m and an inner diameter about four mm. The box was connected to an intake pipe and an outlet pipe with a length of about 20 m. To ensure the gas in the box could be quickly mixed and evenly distributed, two small fans (10 cm in diameter) were set at the top of the box. Each sampling point was measured in a sealed transparent box or dark box for 2 min, and the measured data from the dark box were used to ascertain the CH₄ fluxes. The drought treatment started on July 15, 2017, and end on August 16, 2017. The measurements were taken at three periods of one day (first: 9:00–10:00, second: 12:00–13:00, third: 14:00–15:00). For the measurement of aboveground biomass, 50 cm ×50 cm quadrats were randomly chosen in each experimental plot, and all plants within the quadrats were cut to ground level. After the dust was removed, the plant material was oven dried to constant weight at 70 °C. Belowground biomass was collected by digging soil pits at the same locations where the aboveground biomass had been removed at the sampling depths of 0–20 cm and 20–40 cm. Soils containing root biomass were placed in 40-mesh nylon bags and taken back to the laboratory, where the roots were carefully washed and then oven dried to a constant weight at 70 °C. A soil drill was used to sample the soil via multi-point sampling and mixing. The soil organic matter (SOC) was determined using a potassium dichromate volumetric method (Wang et al., 2007), total carbon (TC) was determined by the elemental analyzer (Sokolova & Vorozhtsov, 2014), and total nitrogen (TN) was determined via the Kjeldahl method (Nozawa et al., 2005).

Data analysis

The formula used for calculating the greenhouse gas fluxes (Wickland et al., 2001) was: (1)${\displaystyle {\mathrm{F}}_{\mathrm{c}}=\frac{\partial {\mathrm{C}}^{\prime }}{\partial \mathrm{t}}\times \frac{\mathrm{M}}{{\mathrm{V}}_0}\times \frac{\mathrm{P}}{{\mathrm{P}}_0}\times \frac{{\mathrm{T}}_0}{{\mathrm{T}}_0+\mathrm{t}}\times \frac{\mathrm{H}}{100}\times 3600}$

where Fc is the gas fluxes (mg C/ (m2 h)); M is the molar mass of gas (g/mol); V₀ is the standard molar volume of gas (22.4 L/mol); P/P₀ is the measurement of pressure to standard air pressure; T₀ is the absolute temperature (273.15 °C); t is the average value of the measured temperature in the box (°C); and H is the static height (cm). Importantly, the measured data were analyzed by linear regression to calculate the linear slope of the gas concentration relative to the time of observation.

Repeated-measure ANOVA with Duncan’s multiple-range tests were performed to examine the main and interaction effects of date, treatment and block on the differences in CH₄ fluxes and environmental factors in 2017 (SPSS, Chicago, IL, USA). A one-way ANOVA analysis was performed to examine the properties (above and below ground biomass, TC, TN, SOC) at different depths in 2017 (SPSS, Chicago, IL, USA). To further evaluate the relationship of CH₄ fluxes and environmental factors, a correlation matrix analysis between CH₄ fluxes and environmental factors was conducted (Origin 2017, USA). The slopes of those linear relationships were analyzed and compared by SMA (Standardized Major Axis) regression analysis, using the SMATR (Standardized Major Axis Tests and Routines) package (Warton et al., 2006). R v3.5.1 with the corrplot package was used for the correlation analysis (Svetnik et al., 2004).

Climate during the experiment period

During the experiment period (32 d), 14 precipitation events occurred in Zoige, with 6 days including ecologically effective precipitation events (≥3 mm). The daily precipitation ranged from 0.1 mm to 20.6 mm (Fig. 2) and the average precipitation was 1.9 mm. The total precipitation was 58.9 mm in the control treatment and 0 mm in the extreme drought treatment during the experimental period. The precipitation mainly occurred in early August, and a transient rainfall occurred at the end of the treatment period. The highest and lowest daily temperatures were 15.3 °C and 8.4 °C, respectively, and the average temperature was 12.9 °C during the treatment period.

Effects of extreme drought on CH₄ fluxes

From the end of June to the middle of July, there was a transition period between a weak CH₄ sink and a weak CH₄ source of the Zoige peatland (Fig. 3A). The emission of CH₄ from the Zoige peatland reached a maximum around August 16. During the pre-drought period, the ecosystem functioned as a CH₄ sink, and during the extreme drought and post-drought periods, the ecosystem functioned as a net CH₄ source (Fig. 3B). Compared to the control treatment, extreme drought significantly decreased the CH₄ fluxes of the Zoige peatland ecosystem by 31.54% (P < 0.05, Fig. 3B, Table 1) in the drought period, and there was no significant change in the pre and post-drought periods of the experiment under the extreme drought and control treatment (P > 0.05). Additionally, the difference of CH₄ fluxes between the control and drought reached the highest value at the peak of plant growth (Fig. 3C). The extreme drought significantly decreased SWC at depths of 5, 10, and 20 cm (P < 0.05), but there was no significant influence of the extreme drought on Ts at depths of 5, 10, or 20 cm (P > 0.05, Table 1).

Effects of extreme drought on plant biomass and soil physicochemical properties

The extreme drought treatment significantly decreased the aboveground biomass of the Zoige alpine peatland ecosystem by 42.75% (P < 0.05, Fig. 4A). The extreme drought treatment significantly decreased the belowground biomass by 59.73% and 59.65% at a depth of 0–10 cm and 10–20 cm, respectively (P < 0.05, Fig. 4B). Under both treatments, the root mass of the subsoil (10–20 cm) was higher than that of the topsoil (0–10 cm) (Fig. 4B). Subsoil (20 cm) SWC was higher than that of the topsoil (5, 10 cm) (Fig. 4C). Significant differences in TC and TN between the two treatments were observed at a depth of 10–20 cm (P < 0.05, Figs. 4D–4E), but there was no significant difference in TC or TN between the extreme drought treatment and control treatment at depths of 0–10 cm (P > 0.05, Figs. 4D–4E). There was also no significant difference in SOC at depths of 0–10 and 10–20 cm (P > 0.05, Fig. 4F). The organic matter (TC, TN, and SOC) of the subsoil was lower than that of the top soil (Figs. 4D–4F).

Relationship between CH₄ fluxes and environmental factors

The regression analysis showed that the Ts at the depth of 10 and 20 cm had a significantly negatively relationship with CH₄ fluxes between the two treatments (P < 0.05, Figs. 5B–5C), as the CH₄ fluxes gradually decreased as the Ts increased. The correlation between the subsoil (20 cm) temperature and CH₄ fluxes was higher than it was with the topsoil (5, 10 cm) temperature between the two treatments (Figs. 5A–5C). The dynamics of the CH₄ fluxes correlated well with that of the SWC, both in the extreme drought and control treatments (Figs. 5D–5F). The SWC at depths of 5 (P < 0.01), 10 (P < 0.01), and 20 (P < 0.01) cm was negatively correlated with the CH₄ fluxes under the extreme drought and control treatments (Figs. 5D–5F). The correlation between the subsoil (20 cm) water content and CH₄ fluxes was higher than it was with the topsoil (5, 10 cm) water content between the two treatments. Moreover, there was a significant difference in the slopes of the SWC at depths of 5, 10, and 20 cm between the control and drought treatments (P_slope < 0.01, Figs. 5D–5F). The slope of the CH₄ fluxes under the extreme drought treatment was lower than that under the control treatment relative to the SWC. The correlation of CH₄ fluxes to SWC was higher than it was relative to Ts (Figs. 5A–5F).

The correlation matrix analysis between CH₄ fluxes and the different environmental factors at depths of 5, 10, and 20 cm were negative under the two treatments. The correlation between CH₄ fluxes and subsoil (20 cm) environmental factors (SWC and Ts) was higher than that with the topsoil (5, 10 cm) environmental factors (Figs. 6A–6D). The extreme drought decreased the correlation between the Ts and CH₄ fluxes (Figs. 6A–6B), and the extreme drought decreased the correlation between the SWC and CH₄ fluxes (Figs. 6C–6D). There was a stronger relationship between the SWC and CH₄ fluxes than between the Ts and CH₄ fluxes (Figs. 6A–6D).